Study 2 - Data Analysis
Veronica Diveica
1 R Session Information
## R version 4.1.1 (2021-08-10)
## Platform: x86_64-apple-darwin17.0 (64-bit)
## Running under: macOS Big Sur 10.16
##
## Matrix products: default
## BLAS: /Library/Frameworks/R.framework/Versions/4.1/Resources/lib/libRblas.0.dylib
## LAPACK: /Library/Frameworks/R.framework/Versions/4.1/Resources/lib/libRlapack.dylib
##
## locale:
## [1] en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8
##
## attached base packages:
## [1] stats graphics grDevices utils datasets methods base
##
## other attached packages:
## [1] bayestestR_0.13.1 rstanarm_2.21.4 Rcpp_1.0.10 ggprism_1.0.4
## [5] interactions_1.1.5 afex_1.3-0 lmerTest_3.1-3 lme4_1.1-33
## [9] Matrix_1.3-4 sjPlot_2.8.14 lubridate_1.9.2 forcats_1.0.0
## [13] stringr_1.5.0 dplyr_1.1.2 purrr_1.0.1 readr_2.1.4
## [17] tidyr_1.3.0 tibble_3.2.1 ggplot2_3.4.2 tidyverse_2.0.0
## [21] plyr_1.8.8
##
## loaded via a namespace (and not attached):
## [1] backports_1.4.1 jtools_2.2.1 igraph_1.4.2
## [4] splines_4.1.1 crosstalk_1.2.0 TH.data_1.1-2
## [7] rstantools_2.3.1 inline_0.3.19 digest_0.6.31
## [10] htmltools_0.5.5 fansi_1.0.4 magrittr_2.0.3
## [13] tzdb_0.3.0 modelr_0.1.11 RcppParallel_5.1.7
## [16] matrixStats_0.63.0 vroom_1.6.3 xts_0.13.1
## [19] sandwich_3.0-2 timechange_0.2.0 prettyunits_1.1.1
## [22] colorspace_2.1-0 xfun_0.39 callr_3.7.3
## [25] crayon_1.5.2 jsonlite_1.8.4 survival_3.5-5
## [28] zoo_1.8-12 glue_1.6.2 gtable_0.3.3
## [31] emmeans_1.8.5 sjstats_0.18.2 sjmisc_2.8.9
## [34] car_3.1-2 pkgbuild_1.4.0 rstan_2.21.8
## [37] abind_1.4-5 scales_1.2.1 mvtnorm_1.1-3
## [40] DBI_1.1.3 ggeffects_1.2.2 miniUI_0.1.1.1
## [43] xtable_1.8-4 performance_0.10.4 bit_4.0.5
## [46] stats4_4.1.1 StanHeaders_2.21.0-7 DT_0.28
## [49] htmlwidgets_1.6.2 threejs_0.3.3 ellipsis_0.3.2
## [52] pkgconfig_2.0.3 loo_2.6.0 sass_0.4.5
## [55] utf8_1.2.3 tidyselect_1.2.0 rlang_1.1.1
## [58] reshape2_1.4.4 later_1.3.1 munsell_0.5.0
## [61] tools_4.1.1 cachem_1.0.8 cli_3.6.1
## [64] generics_0.1.3 sjlabelled_1.2.0 broom_1.0.4
## [67] evaluate_0.20 fastmap_1.1.1 yaml_2.3.7
## [70] processx_3.8.1 knitr_1.42 bit64_4.0.5
## [73] pander_0.6.5 nlme_3.1-162 mime_0.12
## [76] compiler_4.1.1 bayesplot_1.10.0 shinythemes_1.2.0
## [79] rstudioapi_0.14 bslib_0.4.2 stringi_1.7.12
## [82] ps_1.7.5 lattice_0.21-8 nloptr_2.0.3
## [85] markdown_1.6 shinyjs_2.1.0 vctrs_0.6.2
## [88] pillar_1.9.0 lifecycle_1.0.3 jquerylib_0.1.4
## [91] estimability_1.4.1 insight_0.19.2 httpuv_1.6.11
## [94] R6_2.5.1 promises_1.2.0.1 gridExtra_2.3
## [97] codetools_0.2-19 boot_1.3-28.1 colourpicker_1.2.0
## [100] MASS_7.3-60 gtools_3.9.4 withr_2.5.0
## [103] shinystan_2.6.0 multcomp_1.4-23 parallel_4.1.1
## [106] hms_1.1.3 grid_4.1.1 coda_0.19-4
## [109] minqa_1.2.5 rmarkdown_2.21 carData_3.0-5
## [112] numDeriv_2016.8-1.1 shiny_1.7.4 base64enc_0.1-3
## [115] dygraphs_1.1.1.62 The effect of socialness on reaction times (log transformed)
2.1 LMM Results
Iterative procedure suggests modelling a random intercept per participant. Model with optimal random effects structure: logRT ~ Socialness*Concreteness + (1|participant) + (1|Word)
| Â | logRT | ||||
|---|---|---|---|---|---|
| Predictors | Estimates | CI | Statistic | p | df |
| (Intercept) | -0.03 | -0.05 – -0.01 | -2.62 | 0.011 | 83.34 |
| Socialness1 | -0.01 | -0.02 – 0.00 | -1.33 | 0.185 | 123.78 |
| Concreteness | 0.00 | -0.00 – 0.01 | 0.63 | 0.533 | 124.78 |
|
Socialness1 × Concreteness |
-0.00 | -0.02 – 0.01 | -0.68 | 0.496 | 124.76 |
| Random Effects | |||||
| σ2 | 0.01 | ||||
| Ï„00 Word | 0.00 | ||||
| Ï„00 participant | 0.01 | ||||
| ICC | 0.42 | ||||
| Marginal R2 / Conditional R2 | 0.001 / 0.418 | ||||
2.1.1 Check Assumptions
2.2 Bayesian ROPE analysis
The Bayesian analyses estimated that 82.91% of the HDI for socialness (90.4% PD), 100% for concreteness (72.99% PD), and 96.87% for the interaction (75.1% PD), fell within the ROPE (-0.015-0.015).
3 The effect of socialness on accuracy
3.1 LMM
Iterative procedure suggests modelling a random intercept per participant. Model with optimal random effects structure: Accuracy ~ Socialness*Concreteness + (1|participant) + (1|Word)
| Â | ACC | |||
|---|---|---|---|---|
| Predictors | Odds Ratios | CI | Statistic | p |
| (Intercept) | 32.42 | 20.78 – 50.59 | 15.32 | <0.001 |
| Socialness1 | 1.13 | 0.78 – 1.64 | 0.65 | 0.516 |
| Concreteness | 0.94 | 0.78 – 1.13 | -0.70 | 0.483 |
|
Socialness1 × Concreteness |
1.06 | 0.73 – 1.53 | 0.31 | 0.756 |
| Random Effects | ||||
| σ2 | 3.29 | |||
| Ï„00 Word | 0.92 | |||
| Ï„00 participant | 2.80 | |||
| ICC | 0.53 | |||
| Marginal R2 / Conditional R2 | 0.001 / 0.531 | |||
## $Socialness
##
## $Concreteness
3.1.1 Check Assumptions
## $Word
##
## $participant
3.2 ROPE Analysis
The Bayesian analyses estimated that 57.96% of the HDI for socialness (73.86% PD), 90.39% for concreteness (75.37% PD), and 66.15% for the interaction (60.92% PD), fell within the ROPE (-0.181-0.181).
4 The effect of socialness on raw RTs
4.1 LMM
Iterative procedure suggests modelling a random slope per participant. Model with optimal random effects structure: RT ~ Socialness*Concreteness + (1|participant) + (1|Word)
| Â | RT | ||
|---|---|---|---|
| Predictors | Estimates | CI | p |
| (Intercept) | 1.00 | 0.94 – 1.06 | <0.001 |
| Socialness1 | -0.02 | -0.77 – 0.72 | 0.457 |
| Concreteness | 0.00 | -0.01 – 0.02 | 0.613 |
|
Socialness1 × Concreteness |
-0.01 | -0.05 – 0.02 | 0.475 |
| Random Effects | |||
| σ2 | 0.11 | ||
| Ï„00 Word | 0.01 | ||
| Ï„00 participant | 0.06 | ||
| ICC | 0.37 | ||
| N participant | 73 | ||
| N Word | 132 | ||
| Observations | 8621 | ||
| Marginal R2 / Conditional R2 | 0.001 / 0.373 | ||
## $Socialness
##
## $Concreteness
4.1.1 Check Assumptions
4.2 ROPE Analysis
